SQream Advances Big Data Analytics with its GPU-Enabled In-Database Model Training System

SQream, the data and analytics acceleration platform, announced today that it has launched an ‘in-database model training’ feature, to enable use of their solution by customers as both an integrated analytics platform as well as a machine learning model trainer. This expanded offering allows users to enjoy the numerous benefits of in-database model training, including minimized time to insight, faster ingestion, and preparation of large-scale datasets, enhanced model accuracy and precision—and ultimately—more valuable insights. Using the power of SQL accelerated by GPUs to train models directly within the database, SQream maximizes efficiency in machine learning operations and frees up organizations’ valuable resources.

Model training is a critical step that can greatly influence the accuracy and precision of AI-driven predictions. The accuracy of any model is impacted by the quantity and quality of both the training dataset and training algorithm, with the diversity of inputs greatly affecting the accuracy of outputs. SQream’s solution, in contrast with most market offerings that deal with ML use cases, does not require customers to export prepared datasets into a different platform in order to reduce the need to increase predictive accuracy and reduce processing time.

SQream’s in-database model training feature offers a multifaceted value proposition for data-driven organizations employing AI/ML workloads. This feature will be initially released under private preview.The performance advantages are substantial, as the system ensures swift and efficient processing by eliminating the RAM bottleneck during batch model training and minimizes network latency. Cost-effectiveness and return on investment are also prominent, as the in-database training requires fewer compute servers and licenses for designated ML tools. By training models on GPUs, SQream not only enhances ML quality but also provides access to the datasets themselves, significantly improving accuracy and precision as well as broader algorithm support. Furthermore, the solution prioritizes security by ensuring that data never leaves the database, so the integrity of user information will not be compromised.